Just a minor change, so that it's clear that we are setting the supplied
toolchain as the VCPKG_CHAINLOAD_TOOLCHAIN_FILE.
Change-Id: I10b35f04e7ad684015062066f67117ce4a964c1c
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
In order to use multiple toolchain, e.g., when we are cross compiling,
we need to load the detect, and load vcpkg's toolchain after our
toolchain file.
Change-Id: I7f3374363f725a634356788e22ac21835505c4b0
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The method initialized the tab position with the enum value OnlyOneTab.
=> Change this to the correct initial enum value Moving.
Fixes: QTBUG-115147
Pick-to: 6.6
Change-Id: I4ce04f0a41dac6e93affd300eb424f4087eb7867
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
If a style sheet is active, then the palette in the QStyleOption might
be a generated, short-lived, temporary QPalette instance. As QPalette's
cacheKey is based on instance-counters of the private data structures,
it will always be unique for such a QPalette, even if the brushes are
the same as in a previous instance.
(QPalette::cacheKey is just a 64bit integer, we cannot possibly encode
the entire QPalette data in it in a collision-free way, and since a
brush in the palette might contain a pixmap or a gradient we couldn't
even generate an efficient string representation for it. We could at
most cache and reuse QPalette instances based on the attributes in the
style sheet rule and in the base palette of the widget. However, this
seems fragile; it might be an opportunity for future optimization.)
Some styles use the QPixmapCache, with a key that includes the
palette's cache key. The key will always be unique if the palette is
based on style sheet rules, and then we fill pixmap cache with pixmaps
that can never be reused, making the cache both useless and wasteful.
To solve this, generate an empty key if we detect that it is for a style
object that is the target of a style sheet. Return an empty cache key
string from QStyleHelper::uniqueName, which will make QPixmapCache
return immediatey when trying to insert or find an object.
This is not pretty, but it makes the change minimal and low-risk.
Refactoring the respective code paths to e.g. consistently use the
BEGIN_STYLE_PIXMAPCACHE helper macro requires larger changes that
can only be verified visually, and so are out of scope for a bug fix.
This requires changes to code that uses QStyleHelper::uniqueName, as we
need to avoid that other key elements are appended to the generated (and
maybe empty) key. As a side effect, this ends up with code that makes
better use of QStringBuilder.
Pick-to: 6.6
Fixes: QTBUG-114473
Change-Id: I011aed0885f105cbf1e8c0bc6b94c46df47761a3
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This commit provides a minimal API for capturing a graphics
frame, save it and replay it later for debugging. The intention
here is to provide the basic need for future work to allow capturing
through tooling or programmatically from code. This API is
intended to be cross-platform by using Metal Capture Manager on
Apple devices and RenderDoc C++ API everywhere else.
Task-number: QTBUG-114067
Change-Id: If72d92bdef5e5985a0ec2e85e97fd1182da3c53c
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Correct the typo inherts to inherits in QWidget::backgroundRole() docs,
as it is incorrect english.
Fixes: QTBUG-115160
Change-Id: I03892753d6ec7d81c1f818fd9caa279ca11cc3ec
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
ASSERT: "size_t(i) < size_t(size())" in file qbytearray.h, line 492
due to info being emptied out completely and then the code does
while ((info.at(0) == '*')
info was empty because the recent fix "that wasn't the function argument
list" would exit the loop with pos at end.
Incidentally, this change fixes the fact that qCleanupFuncInfo was
removing lambdas:
main(int, char**)::<lambda()>
became
main(int, char**)::
which was, well, shorted, but weird.
Pick-to: 6.6 6.5
Change-Id: Ic7e8f21ea0df7ef96a3f25c4136a727dc0def207
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Amend 3e59a88e89, which incorrectly used
isHidden() to test whether the scrollbar is visible or not.
QWidget::isHidden() is only true for child widgets that are explicitly
hidden (or created for visible parents, which the scrollbars are not).
Since the scrollbars are children of a container that is hidden and
shown, isHidden always returns false.
Instead, use QWidget::isVisibleTo, passing the scroll area, as that
tells us if the scrollbar's visibility is relevant for the layout of the
scroll area.
Add a test case for QAbstractScrollArea, verifying that the scrollbar's
size is correctly taken into account when calculating the size hint.
This change revealed an instability in the tests introduced in the
earlier commit: the layout process is asynchronous, requiring event
processing to update the visibility of the scrollbars. Add a call to
processEvents before storing the reference size hint. Also, explicitly
set a style that doesn't use transient scrollbars as otherwise we cannot
control when the scrollbars are shown.
The chagne also revealed an inaccuracy in the QListView test, which
only passed because the width of the vertical scrollbar was included.
We cannot use font metrics results to compare expected width, as the
item delegate's text rendering uses text layouts.
Task-number: QTBUG-69120
Fixes: QTBUG-109326
Fixes: QTBUG-113552
Pick-to: 6.6 6.5
Change-Id: I1f06f9e88046a77722291ac17c56090f8dff7cf3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
QXmlStreamReader accepted multiple DOCTYPE elements, containing DTD
fragments in the XML prolog, and in the XML body.
Well-formed but invalid XML files - with multiple DTD fragments in
prolog and body, combined with recursive entity expansions - have
caused infinite loops in QXmlStreamReader.
This patch implements a token check in QXmlStreamReader.
A stream is allowed to start with an XML prolog. StartDocument
and DOCTYPE elements are only allowed in this prolog, which
may also contain ProcessingInstruction and Comment elements.
As soon as anything else is seen, the prolog ends.
After that, the prolog-specific elements are treated as unexpected.
Furthermore, the prolog can contain at most one DOCTYPE element.
Update the documentation to reflect the new behavior.
Add an autotest that checks the new error cases are correctly detected,
and no error is raised for legitimate input.
The original OSS-Fuzz files (see bug reports) are not included in this
patch for file size reasons. They have been tested manually. Each of
them has more than one DOCTYPE element, causing infinite loops in
recursive entity expansions. The newly implemented functionality
detects those invalid DTD fragments. By raising an error, it aborts
stream reading before an infinite loop occurs.
Thanks to OSS-Fuzz for finding this.
Fixes: QTBUG-92113
Fixes: QTBUG-95188
Pick-to: 6.6 6.5 6.2 5.15
Change-Id: I0a082b9188b2eee50b396c4d5b1c9e1fd237bbdd
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
It's a bit cumbersome, but works, in principle, using CTAD.
Pick-to: 6.6 6.5 6.2
Task-number: QTBUG-114200
Change-Id: Ib7354180e870a695a978edabf684aedfcf9d9ecc
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Added a template that we can use instead of rewriting the message for
every policy.
Pick-to: 6.6
Change-Id: I13cc182244d5f092e3d5677664bc149c6b126da5
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This reverts commit c2a92199b57b195176d2a0d68d140d72c1cbfb71
"QLibrary::setFileNameAndVersion: reset the tag after findOrCreate".
This restores the behavior of resolve() and compatibility with Qt 4 and
5, which is documented to imply a call to load().
Do note that if you call load() or resolve() and don't call unload(),
the library you've loaded can never be unloaded now. So don't leak!
[ChangeLog][Important Behavior Changes] QLibrary::isLoaded() now reports
whether this instance of QLibrary has succeeded in loading the library,
via direct or indirect call to load(). Previously, it used to reported
whether the actual library was loaded by any QLibrary instance.
The change to QLibrary::resolve() itself is effectively a no-op in this
patch, because isLoaded() would have returned false, but it ensures that
the implementation does what it says it will do.
Fixes: QTBUG-114977
Pick-to: 6.6
Change-Id: I907aa7aea8ef48469498fffd176d7a76ae73e04a
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This avoids us committing to the QByteArray return value, which is
overkill for short strings. Instead, pull the streaming of the unit
behind the ABI boundary, so we're free to change the implementation to
either stream directly or use SSO'ed std::string.
Pick-to: 6.6
Change-Id: I10927acb9d64077d9018b667958ca16be218012a
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
qt_internal_add_tool has the implicit logic that disregards the
previously installed flags and adds new flags that overrides the
exceptions behavior. Add the tool-specific option to make sure
that this won't happen.
Ammends 0f5fbe369b
Pick-to: 6.6 6.5
Change-Id: Id149d8d8002390581813475124ba37dd39d56b9b
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We are gradually enabling more tests for WebAssembly platform
for better test coverage.
Long linking time is no longer an issue due to test batching.
Change-Id: I7ee9f877ecda726bc23d8dd2507c616bb381ebc1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
Change type of mode argument to QDBus::CallMode instead of int.
Also change the argument name in the implementation to much the
declaration.
Change-Id: Ice19bfff72f3bf1f4a5748f2caa16999e87a131f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
It looks cleaner when it's declared where it's defined.
Change-Id: I5006fc086c73e6d4891bc64ff3a8c6b4b17623fe
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Thinking back I'm 99% certain I deliberately used QBA because I was
going to cherry-pick it back and didn't feel like dealing with failing
builds. Coming up we will have a lot of other changes here so let's
change this as well.
Change-Id: Ifc60ad8ee2a1e5b75ac766b4a8b7e1559f80e942
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Change-Id: Ice7f0f026353f5668a437cc35d9b21cde0a69182
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Change-Id: I4148f94a7d9944eddccaba6dbc956abcd2ad9937
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Change-Id: Ibef8da7349d143433522bbb0fd2ee228f0d58135
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
CMake find modules of third-party libraries may create dependency cycles
in their imported library targets. For example, Conan's Vulkan::Vulkan
is such a candidate.
Prevent an infinite loop when generating .pri files for Qt modules that
link against such targets.
Pick-to: 6.5 6.6
Fixes: QTBUG-95569
Change-Id: I09b4a281930f7c6ce5e8c716fe0ab91e2453a24a
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
UIColor.tintColor is only available on iOS 15.0 or newer. Ammends
commit 69cc7f946a
Pick-to: 6.6
Change-Id: Id82006629ee3c3d2548ea39afde265cdccf2cdab
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
QCache::insert on failure destroys the QPixmapCacheEntry, which calls
releaseKey(), which invalidates the key. That's hard to follow, so add
an assert to make it more explicit and prevent future errors.
Pick-to: 6.6
Change-Id: I9c062ef5a6e34c783c064330c914ce4f9a3f9984
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The test hasn't failed on Windows for a long time, but regularly fails
on macOS in CI, so replace the entry accordingly.
Task-number: QTBUG-115154
Pick-to: 6.6
Change-Id: Ib89d15cb9edafad5dd71f6e3f830d03aaeb16331
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
There was a typo in the code that reads the property.
Pick-to: 6.5 6.6
Change-Id: I6a26899ecc6de800fcfaff6da659a611939753ff
Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
Its single parameter is a pointer that defaults null, so don't bother
passing nullptr, much less 0, as it.
Change-Id: Iec68bf388d848c020359001699c971e12266a335
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Previously only toDateTime() was tested. Adding a test-case for
toTime() provoked adding full testing for both it and toDate(), based
on toDateTime() tests.
Pick-to: 6.6 6.5
Task-number: QTBUG-114909
Change-Id: I5c24b3869b3deefc36a7125133822e8f41cd24ba
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
I'll be adding docs for parts currently lacking them; fix up the
numbering first, to disentangle from "real" changes to docs.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: Ia10f212626bf5ca9fab2b6ba6cf02dbd560a2f02
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Renumber the code fragments to match their order, while adding a
number for the previously undocumented custom method. Add a brief
description of it. Move the createMenus() part up to after it, combine
the createActions() with its (as createActions() is long gone, fused
into it and sharing its snippet number).
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: If0fbcadfa058fc12cbd74ba1897646113bd016b0
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
This will make the description of the type flow more naturally.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: I751bdaf420be7afc9cb4925af4f2a94367840605
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Mate Barany <mate.barany@qt.io>
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
The old one showed an old XBEL file's content, with trolltech URLs.
Update to match the new XBEL file's content.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: I2b3bb7d67c6f96b04208fe4b1f9af10055dc0496
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Kimmo Leppälä <kimmo.leppala@qt.io>
The old screenshot showed an old XBEL file's content, with trolltech
URLs. Update to match the new XBEL file's content.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: Idacc31b7786b1e6ed1425857470b1d37227096cd
Reviewed-by: Kimmo Leppälä <kimmo.leppala@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Most links are now https, several have relocated, one entirely
rebranded. Same example in both QXmlStream and DOM forms.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: Ifbc58dadc834cf51113adb1c82de971a8768ee58
Reviewed-by: Cristian Maureira-Fredes <cristian.maureira-fredes@qt.io>
doc.qt.io was described as Qt 5 documentation; no need to version-limit it.
Pick-to: 6.6 6.5
Task-number: QTBUG-111228
Change-Id: I1c5abe3d805c2b845a74b9fe454c494fa1eff4c7
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
As happens for other widgets
(s. QAccessibleWidget::parentObject), report the app
as accessible parent for item views that don't
have another parent set.
Otherwise, the accessible tree is broken when
there's a top-level item view:
The application has the item view as a child,
but the child does not have any parent set.
Extend a QListView autotest accordingly.
Fixes: QTBUG-115135
Pick-to: 6.6 6.5
Change-Id: Ie06874681180a30fc6248dc98f80c4158d837278
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Each line with a test function must be followed by a line describing
the conditions for ignoring that test function.
From what I can see, this wasn't explicitly stated anywhere.
Fixes: QTBUG-114825
Pick-to: 6.2 6.5 6.6
Change-Id: I1941d6c9a6d56a8374a0c5a99c538cf1eaebd544
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
onExit is called whenever the application exits, i.e.
when the app canvas should no longer be displayed and
the loader/embedder code should take some action.
Emscripten provides two callbacks which can be used
here:
- onExit, called when the app exits (but see EXIT_RUNTIME)
- onAbort, called on abort errors.
These map to the two cases Qt's onExit supports. onExit
is not called when EXIT_RUNTIME is disabled, which means
we don't need the special case for exit code 0.
In addition call onExit on any exception. The second
call to showUi() in html_shell.html is then not needed
any more and we avoid duplicating the UI state handling
in user code.
Update the qtloader_integration test to handle changes in
behavior (we no longer set the error text on exit). Use
emscripten_force_exit() to simulate application exit -
using this function makes Emscripten call onExit even
when EXIT_RUNTIME is disabled.
Pick-to: 6.6
Change-Id: I72b5463c1836e8d5054e594abbd304fbc67032b7
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Implement the main features of the pre Qt 6.6 loader
as adaption layer on top of the new loader.
Pick-to: 6.6
Task-id: QTBUG-115049
Change-Id: Iabe860d3fb0488fd003876508787da3688e0c87b
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
... instead of Private ones, at the cost of having to befriend of all
the lockable classes, because we need access to their d_func()'s.
This simplifies the code, because we don't need the manual QClass to
QClassPrivate mapping (o2p) anymore, we can just use d_func(). This also
paves the way to make QEventLoopLocker almost completely inline and use
a 3-pointer form of QBiPointer, once available, to hide the bit
fiddling. We couldn't make such a change if the class continued to hold
pointers to QClassPrivate's.
Pick-to: 6.6
Task-number: QTBUG-114793
Change-Id: Id300e4d45d6cacabe090a46cd6433c5ead3c8b0c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Much less crowded.
Inline QStorageInfoPrivate::root() to ease the split; the Windows
specific code is one extra line.
Change-Id: Icec6822cf436e2b4aa1b3a04184fbfa40e508078
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
They were new... in commit 1f843ca39e
("Add the new QBasicAtomicXXX implementation - no backends yet") for Qt
5.0, which is what allowed us to use the C++11 atomics. That replaced
the even older implementation from Qt 4.4 where QBasicAtomicInt was a
full class and we didn't support any other integer types. That of course
replaced the even older QBasicAtomic from 4.0, first introduced in 2004
(7 years before C++11).
Change-Id: I53335f845a1345299031fffd176fa251b31bfeea
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This is done as one of the last steps inside QProcess itself, so the
child modifier and all other tasks still run with the parent process'
permissions. On Linux, setting the UID to non-zero will also
automatically clear the effective capabilities(7) set.
This feature is only useful for setuid or setgid applications, so this
commit updates the QCoreApplication::setSetuidAllowed() documentation to
mention the QProcess flag.
Change-Id: I3e3bfef633af4130a03afffd175e940c0668d244
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Doing setsid() and disconnecting from the controlling terminal are, in
addition to resetting the standard file descriptors to /dev/null, a
common task that daemons do. These options allow a QProcess to force a
child to be a daemon.
QProcess ensures that the operations are done in the correct order.
Change-Id: I3e3bfef633af4130a03afffd175e9451d2716d7a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The QProcessunixProcessParameters sometimes fails in CI with an XPASS.
Unclear under what conditions QNX behaves correctly, so accept that it
is unpredictable and only expect a failure when a failure is imminent.
Amends f9c87cfd44.
Change-Id: Icf70861343747e6323c7953a2462b7bbc46549b3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
